drift-monitor

drift-monitor is an agent for Claude Code from ShipWithAI/shipwithai-plugins. It costs 67 tokens per session (700 once invoked), scanned A, original, MIT.

A weekly documentation check that compares the repository’s actual files and code with its declared documentation. SSOT means “single source of truth”: the documents meant to describe the current project.

In plain words
What is it for?
Use it to check `CLAUDE.md`, architecture documents, plugin manifests, READMEs, versions, and other declared project details for drift.
Why use it?
It finds documentation that has fallen out of date, so future coding sessions are less likely to follow incorrect project information.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; mentions CLAUDE.md.

Good fit Use it to check CLAUDE.md, architecture documents, plugin manifests, READMEs, versions, and other declared project details for drift.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/shipwithai/shipwithai-plugins/drift-monitor
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/ShipWithAI/shipwithai-plugins

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for drift-monitor

README.md
[![agentmods](https://agentmods.dev/badge/agents/shipwithai/shipwithai-plugins/drift-monitor/github.svg)](https://agentmods.dev/agents/shipwithai/shipwithai-plugins/drift-monitor)
Your own site
<a href="https://agentmods.dev/agents/shipwithai/shipwithai-plugins/drift-monitor"><img src="https://agentmods.dev/badge/agents/shipwithai/shipwithai-plugins/drift-monitor/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for drift-monitor

Your own site · 80×15
<a href="https://agentmods.dev/agents/shipwithai/shipwithai-plugins/drift-monitor"><img src="https://agentmods.dev/badge/agents/shipwithai/shipwithai-plugins/drift-monitor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 67 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 700 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00067 $0.00700
Opus 5 $0.00034 $0.00350
Sonnet 5 $0.00013 $0.00140
Haiku 4.5 $0.00007 $0.00070

Measured 12d ago against content hash 7ff564bb9866, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

drift-monitor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/drift-monitor.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Drift Monitor

Purpose

Detects when CLAUDE.md, docs/ARCHITECTURE.md, and plugin READMEs fall out of sync with the actual repository state. Prevents stale documentation from misleading future Claude sessions.

Context

Reads on startup:

  • CLAUDE.md — declared project structure and conventions
  • docs/ARCHITECTURE.md — declared architecture and key directories
  • plugins/*/manifest.json — declared skill registry per plugin
  • plugins/*/README.md — plugin-level documentation
  • plugins/*/plugin.json — plugin metadata (name, version)

Steps

Step 1 — Scan actual state

Read the real directory tree:

  • List all plugins/*/ directories
  • For each plugin: list skills/ subdirectories and compare against manifest.json
  • Check plugins/*/README.md and plugins/*/CHANGELOG.md for last-updated markers
  • Check plugins/*/plugin.json version against CHANGELOG.md latest entry

Step 2 — Compare against CLAUDE.md

Check each section of CLAUDE.md for accuracy:

  • Key directories — do the listed paths actually exist?
  • Sensitive areas — are the listed paths still present?
  • Conventions — do named tools/patterns match what's in the repo?
  • Harness config — does Tier/Last updated reflect current state?

Step 3 — Compare against docs/ARCHITECTURE.md

  • Key layers table — do the listed directories exist?
  • Entry points — are manifest.json files where declared?
  • Key directories tree — does it match find plugins/ -maxdepth 3 -type d?

Step 4 — Check plugin SSOT

For each plugin directory:

  • Does README.md list all skills currently in skills/?
  • Does manifest.json list all skills in skills/ (no extras, no missing)?
  • Does CHANGELOG.md have an entry for the current version in plugin.json?

Step 5 — Report

Produce a drift report:

## Drift Report — [DATE]

### ✅ In sync
- [item that matches]

### ⚠️ Drifted
- [file/section]: declared "[X]" but actual state is "[Y]"
  → Suggested fix: [one-line fix]

### ❌ Missing
- [file/section that is declared but does not exist]
  → Suggested fix: [one-line fix]

Read the full file on GitHub · 86 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 12d ago First seen · 86 lines · 67 tokens per session scan A 7ff564bb9866

Subscribe to this mod's changes

drift-monitor is an agent published in the GitHub repository ShipWithAI/shipwithai-plugins (10 stars, last pushed 29d ago), licensed MIT. It adds 67 tokens to every session and 700 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

code-quality-plan-creator-default

Architectural Code Quality Agent (LSP-Powered) - Creates comprehensive architectural improvement plans suitable for loop or swarm executors (/implement-loop, /tasks-loop or /tasks-swarm, /beads-loop or /beads-swarm). Uses Claude Code's built-in LSP for semantic code understanding. This agent thoroughly analyzes code…

GantisStorm/essentials-claude-code · 219 tokens

beads-converter-default

Verbatim plan-to-beads converter using the bd CLI. Copies full implementation code, requirements, and exit criteria directly into each bead. Each bead is 100% self-contained - no plan back-references or external lookups needed.

GantisStorm/essentials-claude-code · 55 tokens

bug-plan-creator-default

Architectural Bug Investigation Agent. Deep investigation with line-by-line code analysis, produces fix plans with exact code changes, regression prevention, and verification criteria. Plans work with any executor (loop or swarm).

GantisStorm/essentials-claude-code · 46 tokens

codemap-creator-default

Generate or update hierarchical code maps using LSP. Two modes: create (full scan from root) and update (re-scan only changed files from git diff, MR, or PR). Maps show directory tree with symbols, signatures, dependencies, and export status. Consumed by /plan-creator for codebase orientation.

GantisStorm/essentials-claude-code · 76 tokens

plan-creator-default

Architectural Planning Agent for Brownfield Development. Creates plans for new features with exact code structures, per-file implementation details, and dependency graphs. Plans work with any executor (loop or swarm). For bugs use bug-plan-creator, for code quality use code-quality-plan-creator.

GantisStorm/essentials-claude-code · 61 tokens

document-creator-default

Generate DEVGUIDE.md architectural documentation using LSP for symbol extraction and pattern analysis. Creates .claude/rules/ files when missing. ONLY creates documentation - does not edit existing docs.

GantisStorm/essentials-claude-code · 44 tokens